PTEN mutation spectrum and genotype-phenotype correlations in Bannayan-Riley-Ruvalcaba syndrome suggest a single entity with Cowden syndrome

Abstract

Germline mutations in the tumour suppressor gene PTEN have been implicated in two hamartoma syndromes that exhibit some clinical overlap, Cowden syndrome (CS) and Bannayan-Riley-Ruvalcaba syndrome (BRR). PTEN maps to 10q23 and encodes a dual specificity phosphatase, a substrate of which is phosphatidylinositol 3,4,5-triphosphate, a phospholipid in the phosphatidylinositol 3-kinase pathway. CS is characterized by multiple hamartomas and an increased risk of benign and malignant disease of the breast, thyroid and central nervous system, whilst the presence of cancer has not been formally documented in BRR. The partial clinical overlap in these two syndromes is exemplified by the hallmark features of BRR: macrocephaly and multiple lipomas, the latter of which occur in a minority of individuals with CS. Additional features observed in BRR, which may also occur in a minority of CS patients, include Hashimoto's thyroiditis, vascular malformations and mental retardation. Pigmented macules of the glans penis, delayed motor development and neonatal or infant onset are noted only in BRR. In this study, constitutive DNA samples from 43 BRR individuals comprising 16 sporadic and 27 familial cases, 11 of which were families with both CS and BRR, were screened for PTEN mutations. Mutations were identified in 26 of 43 (60%) BRR cases. Genotype-phenotype analyses within the BRR group suggested a number of correlations, including the association of PTEN mutation and cancer or breast fibroadenoma in any given CS, BRR or BRR/CS overlap family ( P = 0.014), and, in particular, truncating mutations were associated with the presence of cancer and breast fibroadenoma in a given family ( P = 0.024). Additionally, the presence of lipomas was correlated with the presence of PTEN mutation in BRR patients ( P = 0.028). In contrast to a prior report, no significant difference in mutation status was found in familial versus sporadic cases of BRR ( P = 0.113). Comparisons between BRR and a previously studied group of 37 CS families suggested an increased likelihood of identifying a germline PTEN mutation in families with either CS alone or both CS and BRR when compared with BRR alone ( P = 0.002). Among CS, BRR and BRR/CS overlap families that are PTEN mutation positive, the mutation spectra appear similar. Thus, PTEN mutation-positive CS and BRR may be different presentations of a single syndrome and, hence, both should receive equal attention with respect to cancer surveillance.